Like other epistemic activities, inquiry seems to be governed by norms. Some have argued that one such norm forbids us from believing the answer to a question and inquiring into it at the same time. But another, hither-to neglected norm seems to permit just this sort of cognitive arrangement when we seek to confirm what we currently believe. In this paper, I suggest that both norms are plausible and that the conflict between them constitutes a puzzle. Drawing on the felicity conditions of confirmation requests and the biased interrogatives used to perform them, I argue that the puzzle is genuine. I conclude by considering a response to the puzzle that has implications for the debate regarding the relationship between credences and beliefs.